onnxruntime/js/web/lib/wasm/proxy-wrapper.ts
raoanag 424107a82a
Merge main to WindowsAI (#18122)
### Description
Merge main to WindowsAI



### Motivation and Context
<!-- - Why is this change required? What problem does it solve?
- If it fixes an open issue, please link to the issue here. -->

---------

Signed-off-by: Nash <george.nash@intel.com>
Signed-off-by: Yiming Hu <yiming.hu@amd.com>
Signed-off-by: Liqun Fu <liqfu@microsoft.com>
Co-authored-by: Kaz Nishimura <kazssym@linuxfront.com>
Co-authored-by: Tianlei Wu <tlwu@microsoft.com>
Co-authored-by: Nat Kershaw (MSFT) <nakersha@microsoft.com>
Co-authored-by: Yulong Wang <7679871+fs-eire@users.noreply.github.com>
Co-authored-by: Changming Sun <chasun@microsoft.com>
Co-authored-by: zesongw <zesong.wang@intel.com>
Co-authored-by: Yi Zhang <zhanyi@microsoft.com>
Co-authored-by: Dmitri Smirnov <yuslepukhin@users.noreply.github.com>
Co-authored-by: Yifan Li <109183385+yf711@users.noreply.github.com>
Co-authored-by: simonjub <78098752+simonjub@users.noreply.github.com>
Co-authored-by: PeixuanZuo <94887879+PeixuanZuo@users.noreply.github.com>
Co-authored-by: Adrian Lizarraga <adlizarraga@microsoft.com>
Co-authored-by: Edward Chen <18449977+edgchen1@users.noreply.github.com>
Co-authored-by: Arthur Islamov <arthur@islamov.ai>
Co-authored-by: Jambay Kinley <jambaykinley@microsoft.com>
Co-authored-by: Justin Chu <justinchuby@users.noreply.github.com>
Co-authored-by: Wei-Sheng Chin <wschin@outlook.com>
Co-authored-by: Bowen Bao <bowbao@microsoft.com>
Co-authored-by: Hariharan Seshadri <shariharan91@gmail.com>
Co-authored-by: Numfor Tiapo <numsmt2@gmail.com>
Co-authored-by: Vincent Wang <wangwchpku@outlook.com>
Co-authored-by: Pranav Sharma <prs@microsoft.com>
Co-authored-by: George Nash <george.nash@intel.com>
Co-authored-by: Abhishek Jindal <abjindal@microsoft.com>
Co-authored-by: pengwa <pengwa@microsoft.com>
Co-authored-by: Yiming Hu <woinck@users.noreply.github.com>
Co-authored-by: Jiajia Qin <jiajia.qin@intel.com>
Co-authored-by: Lukas Berbuer <36054362+lukasberbuer@users.noreply.github.com>
Co-authored-by: Wanming Lin <wanming.lin@intel.com>
Co-authored-by: Xavier Dupré <xadupre@users.noreply.github.com>
Co-authored-by: aimilefth <60664743+aimilefth@users.noreply.github.com>
Co-authored-by: Baiju Meswani <bmeswani@microsoft.com>
Co-authored-by: Adam Pocock <adam.pocock@oracle.com>
Co-authored-by: Chi Lo <54722500+chilo-ms@users.noreply.github.com>
Co-authored-by: RandySheriffH <48490400+RandySheriffH@users.noreply.github.com>
Co-authored-by: Randy Shuai <rashuai@microsoft.com>
Co-authored-by: Vadym Stupakov <vadim.stupakov@gmail.com>
Co-authored-by: Jian Chen <cjian@microsoft.com>
Co-authored-by: Brian Lambert <98757707+brian-pieces@users.noreply.github.com>
Co-authored-by: Nicolò Lucchesi <nicolo.lucchesi@gmail.com>
Co-authored-by: liqun Fu <liqfu@microsoft.com>
Co-authored-by: trajep <trajepl@gmail.com>
Co-authored-by: Scott McKay <skottmckay@gmail.com>
Co-authored-by: Mustafa Ateş Uzun <mustafauzun0@gmail.com>
Co-authored-by: MistEO <mistereo@hotmail.com>
Co-authored-by: satyajandhyala <satya.k.jandhyala@gmail.com>
Co-authored-by: shaahji <96227573+shaahji@users.noreply.github.com>
Co-authored-by: Rachel Guo <35738743+YUNQIUGUO@users.noreply.github.com>
Co-authored-by: rachguo <rachguo@rachguos-Mini.attlocal.net>
Co-authored-by: Caroline Zhu <wolfivyaura@gmail.com>
Co-authored-by: Caroline Zhu <carolinezhu@microsoft.com>
Co-authored-by: Guenther Schmuelling <guschmue@microsoft.com>
Co-authored-by: xhcao <xinghua.cao@intel.com>
Co-authored-by: Ella Charlaix <80481427+echarlaix@users.noreply.github.com>
Co-authored-by: Xu Xing <xing.xu@intel.com>
Co-authored-by: Hector Li <hecli@microsoft.com>
Co-authored-by: Ye Wang <52801275+wangyems@users.noreply.github.com>
Co-authored-by: Your Name <you@example.com>
Co-authored-by: Benedikt Hilmes <benedikt.hilmes@rwth-aachen.de>
Co-authored-by: rachguo <rachguo@rachguos-Mac-mini.local>
Co-authored-by: George Wu <jywu@microsoft.com>
Co-authored-by: JiCheng <wejoncy@163.com>
Co-authored-by: Sheil Kumar <smk2007@gmail.com>
Co-authored-by: Sheil Kumar <sheilk@microsoft.com>
Co-authored-by: cloudhan <guangyunhan@microsoft.com>
Co-authored-by: kyoshisuki <143475866+kyoshisuki@users.noreply.github.com>
Co-authored-by: aciddelgado <139922440+aciddelgado@users.noreply.github.com>
Co-authored-by: tlwu@microsoft.com <tlwu@a100.crj0ad2y1kku1j4yxl4sj10o4e.gx.internal.cloudapp.net>
Co-authored-by: Maximilian Müller <44298237+gedoensmax@users.noreply.github.com>
Co-authored-by: Tang, Cheng <souptc@gmail.com>
Co-authored-by: Cheng Tang <chenta@microsoft.com@orttrainingdev9.d32nl1ml4oruzj4qz3bqlggovf.px.internal.cloudapp.net>
Co-authored-by: Cheng Tang <chenta@microsoft.com>
Co-authored-by: Jeff Daily <jeff.daily@amd.com>
Co-authored-by: cloudhan <cloudhan@outlook.com>
Co-authored-by: Yufeng Li <liyufeng1987@gmail.com>
Co-authored-by: Zhang Lei <zhang.huanning@hotmail.com>
Co-authored-by: Dwayne Robinson <fdwr@hotmail.com>
Co-authored-by: Zhipeng Han <zhipeng.han@outlook.com>
Co-authored-by: Thiago Crepaldi <thiago.crepaldi@microsoft.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
Co-authored-by: Patrice Vignola <vignola.patrice@gmail.com>
Co-authored-by: kunal-vaishnavi <115581922+kunal-vaishnavi@users.noreply.github.com>
Co-authored-by: snadampal <87143774+snadampal@users.noreply.github.com>
Co-authored-by: Sumit Agarwal <sumitagarwal330@gmail.com>
Co-authored-by: Ashwini Khade <askhade@microsoft.com>
Co-authored-by: Yang Gu <yang.gu@intel.com>
Co-authored-by: Cheng Tang <chenta@a100.crj0ad2y1kku1j4yxl4sj10o4e.gx.internal.cloudapp.net>
Co-authored-by: mindest <30493312+mindest@users.noreply.github.com>
Co-authored-by: Scott McKay <Scott.McKay@microsoft.com>
Co-authored-by: Xavier Dupre <xadupre@microsoft.com@orttrainingdev9.d32nl1ml4oruzj4qz3bqlggovf.px.internal.cloudapp.net>
Co-authored-by: guyang3532 <62738430+guyang3532@users.noreply.github.com>
Co-authored-by: Carson M <carson@pyke.io>
Co-authored-by: sophies927 <107952697+sophies927@users.noreply.github.com>
2023-10-27 17:08:01 -07:00

274 lines
9.6 KiB
TypeScript

// Copyright (c) Microsoft Corporation. All rights reserved.
// Licensed under the MIT License.
import {Env, env, InferenceSession} from 'onnxruntime-common';
import {OrtWasmMessage, SerializableModeldata, SerializableSessionMetadata, SerializableTensorMetadata, TensorMetadata} from './proxy-messages';
import * as core from './wasm-core-impl';
import {initializeWebAssembly} from './wasm-factory';
const isProxy = (): boolean => !!env.wasm.proxy && typeof document !== 'undefined';
let proxyWorker: Worker|undefined;
let initializing = false;
let initialized = false;
let aborted = false;
// resolve; reject
type PromiseCallbacks<T = void> = [(result: T) => void, (reason: unknown) => void];
let initWasmCallbacks: PromiseCallbacks;
let initOrtCallbacks: PromiseCallbacks;
const createSessionAllocateCallbacks: Array<PromiseCallbacks<SerializableModeldata>> = [];
const createSessionFinalizeCallbacks: Array<PromiseCallbacks<SerializableSessionMetadata>> = [];
const createSessionCallbacks: Array<PromiseCallbacks<SerializableSessionMetadata>> = [];
const releaseSessionCallbacks: Array<PromiseCallbacks<void>> = [];
const runCallbacks: Array<PromiseCallbacks<SerializableTensorMetadata[]>> = [];
const endProfilingCallbacks: Array<PromiseCallbacks<void>> = [];
const isOrtEnvInitializedCallbacks: Array<PromiseCallbacks<boolean>> = [];
const ensureWorker = (): void => {
if (initializing || !initialized || aborted || !proxyWorker) {
throw new Error('worker not ready');
}
};
const onProxyWorkerMessage = (ev: MessageEvent<OrtWasmMessage>): void => {
switch (ev.data.type) {
case 'init-wasm':
initializing = false;
if (ev.data.err) {
aborted = true;
initWasmCallbacks[1](ev.data.err);
} else {
initialized = true;
initWasmCallbacks[0]();
}
break;
case 'init-ort':
if (ev.data.err) {
initOrtCallbacks[1](ev.data.err);
} else {
initOrtCallbacks[0]();
}
break;
case 'create_allocate':
if (ev.data.err) {
createSessionAllocateCallbacks.shift()![1](ev.data.err);
} else {
createSessionAllocateCallbacks.shift()![0](ev.data.out!);
}
break;
case 'create_finalize':
if (ev.data.err) {
createSessionFinalizeCallbacks.shift()![1](ev.data.err);
} else {
createSessionFinalizeCallbacks.shift()![0](ev.data.out!);
}
break;
case 'create':
if (ev.data.err) {
createSessionCallbacks.shift()![1](ev.data.err);
} else {
createSessionCallbacks.shift()![0](ev.data.out!);
}
break;
case 'release':
if (ev.data.err) {
releaseSessionCallbacks.shift()![1](ev.data.err);
} else {
releaseSessionCallbacks.shift()![0]();
}
break;
case 'run':
if (ev.data.err) {
runCallbacks.shift()![1](ev.data.err);
} else {
runCallbacks.shift()![0](ev.data.out!);
}
break;
case 'end-profiling':
if (ev.data.err) {
endProfilingCallbacks.shift()![1](ev.data.err);
} else {
endProfilingCallbacks.shift()![0]();
}
break;
case 'is-ort-env-initialized':
if (ev.data.err) {
isOrtEnvInitializedCallbacks.shift()![1](ev.data.err);
} else {
isOrtEnvInitializedCallbacks.shift()![0](ev.data.out!);
}
break;
default:
}
};
const scriptSrc = typeof document !== 'undefined' ? (document?.currentScript as HTMLScriptElement)?.src : undefined;
export const initializeWebAssemblyInstance = async(): Promise<void> => {
if (!BUILD_DEFS.DISABLE_WASM_PROXY && isProxy()) {
if (initialized) {
return;
}
if (initializing) {
throw new Error('multiple calls to \'initWasm()\' detected.');
}
if (aborted) {
throw new Error('previous call to \'initWasm()\' failed.');
}
initializing = true;
// overwrite wasm filepaths
if (env.wasm.wasmPaths === undefined) {
if (scriptSrc && scriptSrc.indexOf('blob:') !== 0) {
env.wasm.wasmPaths = scriptSrc.substr(0, +(scriptSrc).lastIndexOf('/') + 1);
}
}
return new Promise<void>((resolve, reject) => {
proxyWorker?.terminate();
const workerUrl = URL.createObjectURL(new Blob(
[
// This require() function is handled by esbuild plugin to load file content as string.
// eslint-disable-next-line @typescript-eslint/no-require-imports
require('./proxy-worker/main')
],
{type: 'text/javascript'}));
proxyWorker = new Worker(workerUrl, {name: 'ort-wasm-proxy-worker'});
proxyWorker.onerror = (ev: ErrorEvent) => reject(ev);
proxyWorker.onmessage = onProxyWorkerMessage;
URL.revokeObjectURL(workerUrl);
initWasmCallbacks = [resolve, reject];
const message: OrtWasmMessage = {type: 'init-wasm', in : env.wasm};
proxyWorker.postMessage(message);
});
} else {
return initializeWebAssembly(env.wasm);
}
};
export const initializeRuntime = async(env: Env): Promise<void> => {
if (!BUILD_DEFS.DISABLE_WASM_PROXY && isProxy()) {
ensureWorker();
return new Promise<void>((resolve, reject) => {
initOrtCallbacks = [resolve, reject];
const message: OrtWasmMessage = {type: 'init-ort', in : env};
proxyWorker!.postMessage(message);
});
} else {
await core.initRuntime(env);
}
};
export const createSessionAllocate = async(model: Uint8Array): Promise<SerializableModeldata> => {
if (!BUILD_DEFS.DISABLE_WASM_PROXY && isProxy()) {
ensureWorker();
return new Promise<SerializableModeldata>((resolve, reject) => {
createSessionAllocateCallbacks.push([resolve, reject]);
const message: OrtWasmMessage = {type: 'create_allocate', in : {model}};
proxyWorker!.postMessage(message, [model.buffer]);
});
} else {
return core.createSessionAllocate(model);
}
};
export const createSessionFinalize = async(modeldata: SerializableModeldata, options?: InferenceSession.SessionOptions):
Promise<SerializableSessionMetadata> => {
if (!BUILD_DEFS.DISABLE_WASM_PROXY && isProxy()) {
ensureWorker();
return new Promise<SerializableSessionMetadata>((resolve, reject) => {
createSessionFinalizeCallbacks.push([resolve, reject]);
const message: OrtWasmMessage = {type: 'create_finalize', in : {modeldata, options}};
proxyWorker!.postMessage(message);
});
} else {
return core.createSessionFinalize(modeldata, options);
}
};
export const createSession =
async(model: Uint8Array, options?: InferenceSession.SessionOptions): Promise<SerializableSessionMetadata> => {
if (!BUILD_DEFS.DISABLE_WASM_PROXY && isProxy()) {
// check unsupported options
if (options?.preferredOutputLocation) {
throw new Error('session option "preferredOutputLocation" is not supported for proxy.');
}
ensureWorker();
return new Promise<SerializableSessionMetadata>((resolve, reject) => {
createSessionCallbacks.push([resolve, reject]);
const message: OrtWasmMessage = {type: 'create', in : {model, options}};
proxyWorker!.postMessage(message, [model.buffer]);
});
} else {
return core.createSession(model, options);
}
};
export const releaseSession = async(sessionId: number): Promise<void> => {
if (!BUILD_DEFS.DISABLE_WASM_PROXY && isProxy()) {
ensureWorker();
return new Promise<void>((resolve, reject) => {
releaseSessionCallbacks.push([resolve, reject]);
const message: OrtWasmMessage = {type: 'release', in : sessionId};
proxyWorker!.postMessage(message);
});
} else {
core.releaseSession(sessionId);
}
};
export const run = async(
sessionId: number, inputIndices: number[], inputs: TensorMetadata[], outputIndices: number[],
outputs: Array<TensorMetadata|null>, options: InferenceSession.RunOptions): Promise<TensorMetadata[]> => {
if (!BUILD_DEFS.DISABLE_WASM_PROXY && isProxy()) {
// check inputs location
if (inputs.some(t => t[3] !== 'cpu')) {
throw new Error('input tensor on GPU is not supported for proxy.');
}
// check outputs location
if (outputs.some(t => t)) {
throw new Error('pre-allocated output tensor is not supported for proxy.');
}
ensureWorker();
return new Promise<SerializableTensorMetadata[]>((resolve, reject) => {
runCallbacks.push([resolve, reject]);
const serializableInputs = inputs as SerializableTensorMetadata[]; // every input is on CPU.
const message: OrtWasmMessage =
{type: 'run', in : {sessionId, inputIndices, inputs: serializableInputs, outputIndices, options}};
proxyWorker!.postMessage(message, core.extractTransferableBuffers(serializableInputs));
});
} else {
return core.run(sessionId, inputIndices, inputs, outputIndices, outputs, options);
}
};
export const endProfiling = async(sessionId: number): Promise<void> => {
if (!BUILD_DEFS.DISABLE_WASM_PROXY && isProxy()) {
ensureWorker();
return new Promise<void>((resolve, reject) => {
endProfilingCallbacks.push([resolve, reject]);
const message: OrtWasmMessage = {type: 'end-profiling', in : sessionId};
proxyWorker!.postMessage(message);
});
} else {
core.endProfiling(sessionId);
}
};
export const isOrtEnvInitialized = async(): Promise<boolean> => {
if (!BUILD_DEFS.DISABLE_WASM_PROXY && isProxy()) {
ensureWorker();
return new Promise<boolean>((resolve, reject) => {
isOrtEnvInitializedCallbacks.push([resolve, reject]);
const message: OrtWasmMessage = {type: 'is-ort-env-initialized'};
proxyWorker!.postMessage(message);
});
} else {
return core.isOrtEnvInitialized();
}
};